AWRA OpsHub Search

An Imprest for a Till, and None for a Box

This product has a full imprest control — a float, recorded movements, an expected balance, a count and a variance. It exists in exactly one place: the till. The cash box in the back office, which nobody watches, has nothing.

Point of Sale AWRA OpsHub Team 11 min read

Imprest is a two-hundred-year-old idea and it is still the best cash control anybody has invented: fix a float, spend from it, evidence every payment, and top it back up to the float by exactly what you spent. The reconciliation is automatic, because the sum of the cash and the vouchers must always equal the float.

The position, stated first

The imprest pattern is implemented here, correctly, for the cash drawer at a counter. It is absent for petty cash, for field advances and for any float that is not a till. That is backwards relative to where the risk sits, and it is worth knowing before you assume a cash module covers cash.

What the till has

A cash session opens with a float, against a named counter, by a named person. Cash sales accumulate against it. Drops are recorded when the drawer is emptied mid-shift. At close, an expected figure is computed from the float plus sales minus drops, a counted figure is entered, and the variance is stored with a note.

That is an imprest control. Every element is there. It has a supervisor, a rhythm, a documented variance and an audit trail of who opened and who closed.

What the cash box has

Nothing. There is no float record, no voucher, no top-up cycle and no box reconciliation anywhere in this product. A search of the models for anything petty, imprest, voucher or float-shaped returns no result at all.

Petty cash spending is recorded, if it is recorded, as an expense — which posts to the ledger correctly and tells you nothing about whether the box balances.

The same control, two places it could live

The till drawer

Fully controlled

  • An opening float, per session
  • Movements recorded as they happen
  • An expected balance computed continuously
  • A count, a variance and a note
  • Named opener and named closer
  • Supervised, in public, on camera

The office cash box

Not modelled at all

  • No float record
  • No voucher entity
  • No top-up cycle
  • No expected balance to compute
  • No count and no variance
  • Unsupervised, in a drawer, alone

The column with the control is the column with the supervision. That is the wrong way round if you are choosing where to spend a build.

The till is watched by a queue, a camera and a supervisor, and it has an imprest control. The box in the back office is watched by nobody, and it has an expense form.

The third version of the same gap

There is a third consumer of exactly this pattern and it is the one with the sharpest consequences: field advances. Money given to a person to go and do something, accounted for on return, with the unspent balance handed back.

That is imprest again, with a person instead of a box. It does not exist here either — no advance record, no retirement of an advance, no outstanding balance per holder. For organisations that operate on advances rather than on purchase orders, this is consistently the first thing an auditor asks about and the thing they most often find lacking.

  1. Keep the float small and the cycle short

    Weekly top-ups on a small float beat monthly ones on a large float, in every system and in none. The exposure at any moment is the float, not the annual spend.

  2. Record every payment as an expense on the day

    This works today and posts properly to the ledger. It is the evidence half of an imprest without the reconciliation half.

  3. Reconcile the box on paper, on a fixed day

    Cash plus vouchers equals float. It is one line of arithmetic and it does not need software; it needs a fixed day and a second person.

  4. Never let an advance outlive its purpose

    Nothing here will track it, so the control is a list and a date. An advance nobody has retired after ninety days is not an advance.

What we would build

Two, and the pattern is already in the codebase

This is unusually well-defined work because the shape is not in question — the till proves it. What would be built is the same control against a different holder.

A petty cash float with vouchers and top-ups

A float per box with a custodian, payments as vouchers against it, a top-up that restores the float by exactly what was spent, and a reconciliation with a variance and a note. The expense posting already works; what is added is the balance the vouchers have to agree with.

Advances issued to a person, and retired

An advance to a named holder, expenses accounted against it, unspent cash returned and the advance closed — with an outstanding-advances list by holder and by age. This is the one auditors ask for, and the ageing view is most of its value.

How it works: you describe the requirement, we return a written scope, timeline and cost, and once agreed it is built into your environment and maintained as part of the product. If you run on advances rather than on orders, raise this before a trial — it is a structural fit question, not a feature preference.

Talk to us about cash controls

Cash control in this product, precisely

What AWRA OpsHub does today

  • A full imprest-shaped cash session for a till counter: float, drops, expected balance, count, variance, named opener and closer.
  • Expenses recorded and posted to the ledger with double entry, including an approval step where the tenant grants the approve permission.
  • Attachments on expenses, so a receipt can be held against a payment.
  • An expense approval that cannot be granted by whoever raised it, unconditionally and not overridable.

What it does not do

  • A petty cash or imprest module. No float record, no voucher entity, no top-up cycle, no box reconciliation.
  • A field advance or imprest holder. No advance, no retirement, no outstanding balance by person.
  • Any cash control outside the till counter.
  • A cash-in-hand balance anywhere that is not a ledger account.

Not ours, by choice

  • The till control is real and good, and this post is not a criticism of it. The observation is about where it is not, and the asymmetry between supervision and control.
  • Recording petty spending as expenses is not a workaround with a gap in it — it is genuinely half the job, correctly done. The missing half is the balance those payments have to agree with.
  • This post makes no claim about any national statute, and mentions no country. The imprest control is the same everywhere and predates all of them.

Four questions about cash beyond the till

Does the cash module cover anything other than the till?

A good answer sounds like

A specific yes or no.

What it actually means

The question that separates a POS feature from a cash feature. Ours is a no.

Show me the current balance of the office petty cash box.

A good answer sounds like

A figure, on a screen.

What it actually means

If the answer is a ledger account balance, that is accounting, not custody — it will not tell you the box is short.

Who currently holds an unretired advance?

A good answer sounds like

A list, with ages.

What it actually means

The single most useful cash report in an advance-driven organisation, and one of the least often built.

What happens when the box does not balance?

A good answer sounds like

A recorded variance with a reason.

What it actually means

If nothing happens, the box does not really have a control — it has a custodian and a hope.

Ask where the float lives

A cash module that covers the till is a till feature. Ask specifically about the box and the advance, because those are the two places cash goes missing quietly.

Talk about petty cash and advances

Frequently asked questions

Can I use the till cash session for a petty cash box?

Not sensibly. It is bound to a counter and a cashier and it computes its expected balance from till sales, so a box that only pays money out has nothing to reconcile against. The shape is right and the wiring is specific to a till.

Do petty cash payments reach the accounts correctly?

Yes. Expenses post full double entry, and where the tenant has granted the approval permission they also carry an approver who cannot be the person who raised them. The accounting is sound; the custody is not modelled.

Is an advance different from an expense?

Fundamentally. An expense is money spent; an advance is money held. Recording an advance as an expense on the day it is issued overstates spend and loses the balance somebody still owes you, which is precisely the finding auditors write up.

Help Center

Need a quick answer while you read?

Run inventory, procurement, assets, sales, and field work with approved AWRA guidance for setup, migration, integrations, security, pricing, and support.

Search all approved AWRA public help articles.

Open Help Center